Re: [PATCH net-next 2/2] net/smc: introduce statistics for ringbufs usage of net namespace

[Simon Horman <horms@kernel.org>]

On Tue, Aug 06, 2024 at 09:07:40PM +0800, Wen Gu wrote:
> 
> 
> On 2024/8/6 18:49, Simon Horman wrote:
> > On Mon, Aug 05, 2024 at 05:05:51PM +0800, Wen Gu wrote:
> > > The buffer size histograms in smc_stats, namely rx/tx_rmbsize, record
> > > the sizes of ringbufs for all connections that have ever appeared in
> > > the net namespace. They are incremental and we cannot know the actual
> > > ringbufs usage from these. So here introduces statistics for current
> > > ringbufs usage of existing smc connections in the net namespace into
> > > smc_stats, it will be incremented when new connection uses a ringbuf
> > > and decremented when the ringbuf is unused.
> > > 
> > > Signed-off-by: Wen Gu <guwen@linux.alibaba.com>
> > 
> > ...
> > 
> > > diff --git a/net/smc/smc_stats.h b/net/smc/smc_stats.h
> > 
> > ...
> > 
> > > @@ -135,38 +137,45 @@ do { \
> > >   } \
> > >   while (0)
> > > -#define SMC_STAT_RMB_SIZE_SUB(_smc_stats, _tech, k, _len) \
> > > +#define SMC_STAT_RMB_SIZE_SUB(_smc_stats, _tech, k, _is_add, _len) \
> > >   do { \
> > > +	typeof(_is_add) is_a = (_is_add); \
> > >   	typeof(_len) _l = (_len); \
> > >   	typeof(_tech) t = (_tech); \
> > >   	int _pos; \
> > >   	int m = SMC_BUF_MAX - 1; \
> > >   	if (_l <= 0) \
> > >   		break; \
> > > -	_pos = fls((_l - 1) >> 13); \
> > > -	_pos = (_pos <= m) ? _pos : m; \
> > > -	this_cpu_inc((*(_smc_stats)).smc[t].k ## _rmbsize.buf[_pos]); \
> > > +	if (is_a) { \
> > > +		_pos = fls((_l - 1) >> 13); \
> > > +		_pos = (_pos <= m) ? _pos : m; \
> > > +		this_cpu_inc((*(_smc_stats)).smc[t].k ## _rmbsize.buf[_pos]); \
> > > +		this_cpu_add((*(_smc_stats)).smc[t].k ## _rmbuse, _l); \
> > 
> > Nit:
> > 
> > I see that due to the construction of the caller, SMC_STAT_RMB_SIZE(),
> > it will not occur. But checkpatch warns of possible side effects
> > from reuse of _smc_stats.
> > 
> > As great care seems to have been taken in these macros to avoid such
> > problems, even if theoretical, perhaps it is worth doing so here too.
> > 
> > f.e. A macro-local variable could store (*(_smc_stats)).smc[t] which
> >       I think would both resolve the problem mentioned, and make some
> >       lines shorter (and maybe easier to read).
> > 
> 
> It makes sense. I will use a macro-local variable of smc_stats. Thank you!

Great, thanks.